As we enter the era of personalized medicine, characterization of the cancer genome will continue to influence diagnostic and therapeutic decisions in the clinic. Recognizing this, large-scale efforts by The Cancer Genome Atlas (TCGA) and others are generating a compendium of genomic aberrations found across major cancer types with the goal of identifying new therapeutic targets and early detection biomarkers. The challenge now is to find ways to identify functional driver aberrations, as targeting driver events or their activated pathways offers the greatest hope of improving patient outcomes. Oncogenic transcript fusions resulting from chromosomal rearrangements represent an important class of such events, and the successful targeting of fusion oncoproteins such as BCR-ABL1 and EML4-ALK provide strong rationale for comprehensive testing of fusion genes identified in cancer. Unfortunately, the functional interrogation of fusion genes is complicated by the large quantity identified, inability to accurately predict those with driver activity and significant technical roadblocks preventing fusion gene construction for biological assays. To address these challenges, we developed novel methodologies for (1) in silico annotation of NGS data to select novel in-frame gene fusions across diverse tumor types, (2) high-throughput fusion gene construction using a novel recombineering strategy and our platform of >35,000 human open reading frame (ORF) clones, and (3) lentiviral delivery of fusion genes to cell models to identify those with driver activity and responsiveness to available therapeutics. The goal of our project is to scale these technologies for the comprehensive analysis of gene fusions in cancer, ultimately allowing functionalization of thousands of fusion events across diverse cancer types. In Aim 1 we will scale fusion gene construction to model up to 296 fusion genes chosen for their inclusion of druggable protein kinase domains. In Aim 2, fusion genes will be entered into our existing Ba/F3 driver screening platform, which rapidly quantitates the ability of each fusion gene to induce cell survival and proliferation. In Aim 3 we will functionally validate the top fusion drivers identified in Aim 2, followed by analysis of their activity and therapeutic sensitivities by proteomic profiling and compound screening, respectively. Finally, we will subject the top fusions to in vivo validation using engineered context- specific cell models in mice that provide the appropriate genetic and microenvironmental contexts for driver gene validation. This level of technology development, which is widely applicable to all cancer types, will have a sustained impact by creating unique opportunities for transformative research. These systems will reveal the highest priority fusion gene targets to enroll in deep mechanistic biology studies, drug discovery and development programs ultimately leading to personalized treatment strategies.